| axw | wallyworld: gtg take kids to school, can you please babysit https://github.com/juju/juju/pull/6996 and let sinzui know if/when it's landed? | 00:22 |
|---|---|---|
| === frankban is now known as frankban|afk | ||
| wallyworld | axw: wil do, i'll also forward port to develop | 00:23 |
| anastasiamac | PR 6996 and 6998 landed. wallyworld, thumper, menn0, axw, sinzui - rc2 is good to go, right? | 00:33 |
| wallyworld | anastasiamac: yeah, i just let sinzui know in #juju | 00:34 |
| wallyworld | one step pahead of you :-P | 00:34 |
| menn0 | nice | 00:34 |
| * wallyworld merges 2.1 into develop | 00:34 | |
| anastasiamac | wallyworld: one step and probably many years :D | 00:35 |
| wallyworld | :-( :-( :-( | 00:35 |
| wallyworld | about 9 :-( | 00:35 |
| anastasiamac | just keep "experienced" in mind | 00:36 |
| anastasiamac | :D | 00:36 |
| anastasiamac | and hey! I am still somewhere in my 20s \o/ | 00:36 |
| wallyworld | sinzui: wait | 00:37 |
| wallyworld | looks like the merge was done into 2.2 branch | 00:37 |
| wallyworld | let me check | 00:37 |
| wallyworld | sinzui: yeah it was :-( | 00:38 |
| wallyworld | i'll have to quickly backport | 00:38 |
| axw | doh | 01:04 |
| axw | wallyworld: sorry, didn't notice that | 01:05 |
| wallyworld | axw: no worries, backporting now but lots of conflicts to sort out, i'll propose soon, just getting tests to pass | 01:05 |
| wallyworld | axw: i have a failing test in tools/lxdclient. it may be due to the changed socket stuff? i've resolved all the conflicts, diff looks ok, local lxd provider test with IP4 is ok. you may know immediately what issue is? https://github.com/juju/juju/compare/2.1...wallyworld:backport-pr-6996?expand=1 | 01:38 |
| * axw looks | 01:38 | |
| axw | wallyworld: pastebin the failure? | 01:38 |
| wallyworld | axw: http://pastebin.ubuntu.com/24010792/ | 01:39 |
| wallyworld | ipv6 | 01:40 |
| wallyworld | code looks ok to me at first glance | 01:40 |
| wallyworld | there's a bit there in fixAddr to sort out IPv6 | 01:40 |
| axw | wallyworld: fixAddr isn't even used? :/ | 01:41 |
| axw | wallyworld: delete addserver*.go I think | 01:42 |
| wallyworld | axw: huh, yeah, i think you're right | 01:42 |
| wallyworld | hmmmm | 01:42 |
| wallyworld | axw: doesn't appear used in develop branch either | 01:44 |
| axw | wallyworld: chop it. I've never come across it before | 01:44 |
| axw | hasn't been touched since 2015 | 01:45 |
| wallyworld | axw: https://github.com/juju/juju/pull/6999 | 01:47 |
| wallyworld | thumper: free for 1:1? | 01:49 |
| thumper | yeha | 01:49 |
| axw | wallyworld: LGTM, thanks | 01:50 |
| wallyworld | yay | 01:50 |
| wallyworld | axw: talking to tim so will be late for 1:1 | 01:51 |
| axw | wallyworld: np, ping when | 01:57 |
| wallyworld | axw: free now | 02:14 |
| axw | wallyworld: just a minute | 02:14 |
| babbageclunk | thumper: easy review? https://github.com/juju/juju/pull/7000 | 02:30 |
| * thumper looks | 02:37 | |
| menn0 | axw: i'm still looking but goswagger is looking like the best fit so far | 02:38 |
| thumper | babbageclunk: code looks fine, but left a comment for you to check | 02:38 |
| axw | menn0: it would be great if that can be made to work, it's pretty much the de facto standard now | 02:39 |
| babbageclunk | thumper: thanks | 02:39 |
| axw | menn0: are there any existing examples of mapping between RPC-style and REST-style with it? | 02:39 |
| menn0 | axw: for the existing rpc API I'm wondering about either generating it from the same swagger yaml or having an RPC to REST layer | 02:40 |
| axw | menn0: not sure, I have no hands on experience | 02:41 |
| babbageclunk | thumper: yeah, I tried that as well, it worked fine. | 02:42 |
| thumper | ok, lgtm | 02:43 |
| axw | menn0: somewhat related: https://github.com/OAI/OpenAPI-Specification/issues/523 | 02:47 |
| axw | menn0 babbageclunk: did either of you fix some proxy related bugs recently? anything that would affect bootstrap on lxd? | 03:03 |
| menn0 | axw: babbageclunk did make some proxy fixes recently. not sure how they would affect bootstrap though. what's happening? | 03:04 |
| axw | menn0: https://bugs.launchpad.net/juju/+bug/1633788/comments/30 | 03:04 |
| mup | Bug #1633788: juju 2.0.0 bootstrap to lxd fails (connect to wrong "remote" IP address) <canonical-is> <juju> <lxd> <lxd-provider> <s390x> <uosci> <OpenStack Charm Test Infra:Confirmed for 1chb1n> <juju:Fix Released by axwalk> <https://launchpad.net/bugs/1633788> | 03:05 |
| axw | menn0: 2017-02-16 12:00:56 ERROR cmd supercommand.go:458 new environ: Get https://10.87.204.1:8443/1.0: Service Unavailable | 03:05 |
| axw | menn0: AFAICS, lxd won't return 503. so I'm thinking it's going through the proxy | 03:06 |
| menn0 | axw: I don't think the proxy settings would ever have been used during bootstrap b/c the proxyupdater worker doesn't run during bootstrap | 03:06 |
| axw | I just tested, but now I realise I didn't set https_proxy, only http_proxy | 03:07 |
| axw | hmm ok | 03:07 |
| babbageclunk | it should use the http_proxy for https if there's no https_proxy set. | 03:08 |
| babbageclunk | menn0: would the proxy settings be set in the env for jujud after bootstrap? | 03:09 |
| axw | babbageclunk menn0: looks like we do set the proxy settings in cloud-init | 03:10 |
| axw | babbageclunk menn0: so they would also affect the bootstrap command | 03:11 |
| babbageclunk | axw: gah, that'll be it then. | 03:11 |
| babbageclunk | axw: I think you could fix it by initialising juju/juju/utils/proxy.DefaultConfig utils.proxy.DetectProxies() | 03:13 |
| babbageclunk | want me to do a diff for it? | 03:13 |
| axw | babbageclunk: what will that do? | 03:14 |
| babbageclunk | It will set the initial values of the proxy config to whatever's in the env vars. | 03:15 |
| babbageclunk | axw: Then after that any changes can be made from the proxyupdater. | 03:19 |
| axw | babbageclunk: that sounds sane I think | 03:19 |
| axw | babbageclunk: if you would, that'd be good - thanks | 03:20 |
| babbageclunk | axw: what should I do about an error from initialising a package variable? There's probably no logging available and panicking seems nasty. | 03:25 |
| axw | babbageclunk: can we just do it in cmd/jujud/main.go, and error out if it fails? | 03:26 |
| axw | babbageclunk: before we install the proxy thing into the default transport? | 03:26 |
| babbageclunk | axw: yeah, sounds good | 03:28 |
| babbageclunk | axw: I'll make the PR against 2.1 | 03:36 |
| babbageclunk | axw: https://github.com/juju/juju/pull/7002 | 03:42 |
| babbageclunk | axw: trying it out here | 03:42 |
| babbageclunk | axw: Sorry I missed that - because the original problem was about cached env vars, my QA steps were "bootstrap without proxy set up, deploy something, set proxy, deploy something else" | 03:51 |
| babbageclunk | axw: I should've thought to check bootstrapping with proxy | 03:51 |
| axw | babbageclunk: all good | 03:51 |
| babbageclunk | axw: also, menn0 misled me. :) | 03:51 |
| axw | this is why we have RCs :) | 03:51 |
| menn0 | babbageclunk: wat?! :) | 03:52 |
| babbageclunk | menn0: well, I should've checked anyway | 03:54 |
| babbageclunk | axw: so should I ping sinzui so he knows that this should be included in the release? | 03:55 |
| axw | babbageclunk: wouldn't hurt, I'm not sure if it's too late | 03:56 |
| menn0 | babbageclunk, axw: it just occurred to me that what might be confusing to people is that if they bootstrap with --config http_proxy=.... but with $http_proxy not set on the host machine, that proxy on the command line isn't going to get used | 03:56 |
| menn0 | I guess that's ok | 03:57 |
| menn0 | you don't necessary want the client doing the bootstrap to use the same proxy as the juju deployment | 03:57 |
| axw | menn0: I think we need to support that case too. i.e. the client goes direct, but the env goes thru proxy | 03:57 |
| axw | right | 03:57 |
| sinzui | babbageclunk change the proxy config after deploy and expect it to be honoured? | 03:58 |
| babbageclunk | sinzui: oh hai | 03:58 |
| babbageclunk | sinzui: wasn't expecting you to be up | 03:58 |
| sinzui | I was just checking in on the possible last commit | 03:59 |
| babbageclunk | sinzui: axw just pointed out a comment that showed my proxy fix didn't work with bootstrapping. | 04:00 |
| babbageclunk | sinzui: the fix is building now | 04:00 |
| babbageclunk | sinzui: is it too late? | 04:00 |
| sinzui | babbageclunk: building as in not merged? | 04:01 |
| axw | babbageclunk: it's not 100% clear that's the issue there, but it will be in other scenarios at least | 04:01 |
| babbageclunk | axw: true | 04:01 |
| babbageclunk | sinzui: yes | 04:01 |
| babbageclunk | sinzui: https://github.com/juju/juju/pull/7002 | 04:01 |
| sinzui | babbageclunk: :/ | 04:02 |
| babbageclunk | sinzui: sorry :( | 04:02 |
| sinzui | babbageclunk: not too late. | 04:02 |
| sinzui | babbageclunk: The maas I was worked about is still up so I hope all eill be tested when I wake. I expect to need to reest a few bad substrates like a maas or azure | 04:07 |
| babbageclunk | sinzui: it's landed on 2.1 now | 04:33 |
| sinzui | thank you babbageclunk . I ma still reetting a maas so that you branch will pass first time | 04:34 |
| sinzui | though I think I just worked out how the maas 1.9 is breaking. the squid might be omming | 04:35 |
| babbageclunk | the squid might be nomming. On some Patagonian toothfish or something. | 04:49 |
| wallyworld | babbageclunk: your latest PR just missed my forward port of 2.1 into develop | 05:16 |
| babbageclunk | wallyworld: ok - I'll pull it forward into develop now | 06:06 |
| babbageclunk | wallyworld: ping? | 06:39 |
| wallyworld | babbageclunk: hey | 06:42 |
| babbageclunk | wallyworld: sorry - was having trouble with a merge but I think I sorted it - magit was leading me astray. | 06:43 |
| wallyworld | no worries, i was just at school pickup | 06:43 |
| babbageclunk | get anything good? | 06:43 |
| wallyworld | nope, just a surly 15 year old | 06:45 |
| wallyworld | axw: lgtm on your pr | 06:45 |
| axw | wallyworld: ta | 06:46 |
| babbageclunk | wallyworld: actually, I'm still having trouble. How should I be doing the merge? My branch is reporting heaps of weird differences with develop. | 06:46 |
| babbageclunk | wallyworld: https://github.com/juju/juju/compare/develop...babbageclunk:merge-proxy-fix?expand=1 | 06:47 |
| wallyworld | i'd use cheery pick | 06:47 |
| wallyworld | git cherrypick -m 1 <sha> | 06:47 |
| wallyworld | merge will give you grief | 06:47 |
| babbageclunk | wallyworld: ah, ok - that's how I'd normally do it, but I thought jam preferred using a merge | 06:47 |
| babbageclunk | wallyworld: that it is | 06:48 |
| wallyworld | yeah, i merged 2.1 this morning and it was ok | 06:48 |
| babbageclunk | wallyworld: ok, cherry picking I can do | 06:48 |
| wallyworld | for a single commit i tend to cheery pick | 06:48 |
| wallyworld | but that's just me | 06:48 |
| === frankban|afk is now known as frankban | ||
| perrito666 | morning | 11:51 |
| alexisb_ | morning perrito666 | 14:17 |
| perrito666 | alexisb_: morning | 14:19 |
| === tvansteenburgh1 is now known as tvansteenburgh | ||
| jam | babbageclunk: so you can cherry pick, but that makes future merges harder | 16:07 |
| jam | as long as it is 2.1 => develop, I wouldn't try develop => 2.1 | 16:07 |
| === frankban is now known as frankban|afk | ||
| babbageclunk | sorry jam, sounds like I just contributed to the problem again then. Next time I need to do it I'll try to pick your brains. It sounds like I should have just accepted the differences it was reporting? | 19:28 |
| perrito666 | I believe EOD/EOW reached me, I am really tired | 19:40 |
| perrito666 | cheers all | 19:40 |
| === mup_ is now known as mup | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!